This document describes how the system behaves while running.
- Devices publish telemetry periodically
- Heartbeats update device liveness
- Core processes messages asynchronously
- Dashboards reflect near-real-time state
- Missing heartbeats mark a device as degraded
- No retries from backend toward devices
- Device state is preserved until timeout
- Devices reconnect automatically
- Core resubscribes on reconnect
- No manual intervention required
- systemd restarts services
- Core rebuilds in-memory state from persisted data
- Devices reconnect naturally
- Telemetry is best-effort
- Events are prioritized
- Health state is eventually consistent